Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Prediction Algorithms”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 37 records · Page 2Linked to original sources
PubMed2006

Protein-lipid interactions: correlation of a predictive algorithm for lipid-binding sites with three-dimensional structural data.

BACKGROUND: Over the past decade our laboratory has focused on understanding how soluble cytoskeleton-associated proteins interact with membranes and other lipid aggregates. Many protein domains mediating specific cell membrane interactions appear by fluorescence microscopy and other precision techniques to be partially inserted into the lipid bilayer. It is unclear whether these protein-lipid-interactions are dependent on shared protein motifs or unique regional physiochemistry, or are due to more global characteristics of the protein. RESULTS: We have developed a novel computational program that predicts a protein's lipid-binding site(s) from primary sequence data. Hydrophobic labeling, Fourier transform infrared spectroscopy (FTIR), film balance, T-jump, CD spectroscopy and calorimetry experiments confirm that the interfaces predicted for several key cytoskeletal proteins (alpha-actinin, Arp2, CapZ, talin and vinculin) partially insert into lipid aggregates. The validity of these predictions is supported by an analysis of the available three-dimensional structural data. The lipid interfaces predicted by our algorithm generally contain energetically favorable secondary structures (e.g., an amphipathic alpha-helix flanked by a flexible hinge or loop region), are solvent-exposed in the intact protein, and possess favorable local or global electrostatic properties. CONCLUSION: At present, there are few reliable methods to determine the region of a protein that mediates biologically important interactions with lipids or lipid aggregates. Our matrix-based algorithm predicts lipid interaction sites that are consistent with the available biochemical and structural data. To determine whether these sites are indeed correctly identified, and whether use of the algorithm can be safely extended to other classes of proteins, will require further mapping of these sites, including genetic manipulation and/or targeted crystallography.

Algorithms↗
PubMed1999

Differentiating between septic arthritis and transient synovitis of the hip in children: an evidence-based clinical prediction algorithm.

BACKGROUND: A child who has an acutely irritable hip can pose a diagnostic challenge. The purposes of this study were to determine the diagnostic value of presenting variables for differentiating between septic arthritis and transient synovitis of the hip in children and to develop an evidence-based clinical prediction algorithm for this differentiation. METHODS: We retrospectively reviewed the cases of children who were evaluated at a major tertiary-care children's hospital between 1979 and 1996 because of an acutely irritable hip. Diagnoses of true septic arthritis, presumed septic arthritis, and transient synovitis were explicitly defined on the basis of the white blood-cell count in the joint fluid, the results of cultures of joint fluid and blood, and the clinical course. Univariate analysis and multiple logistic regression analysis were used to compare groups. A probability algorithm for differentiation between septic arthritis and transient synovitis on the basis of independent multivariate predictors was constructed and tested. RESULTS: Patients who had septic arthritis differed significantly (p < 0.05) from those who had transient synovitis with regard to the erythrocyte sedimentation rate, serum white blood-cell count and differential, weight-bearing status, history of fever, temperature, evidence of effusion on radiographs, history of chills, history of recent antibiotic use, hematocrit, and gender. Patients who had true septic arthritis differed significantly (p < 0.05) from those who had presumed septic arthritis with regard to history of recent antibiotic use, history of chills, temperature, erythrocyte sedimentation rate, history of fever, gender, and serum white blood-cell differential. Four independent multivariate clinical predictors were identified to differentiate between septic arthritis and transient synovitis: history of fever, non-weight-bearing, erythrocyte sedimentation rate of at least forty millimeters per hour, and serum white blood-cell count of more than 12,000 cells per cubic millimeter (12.0 x 10(9) cells per liter). The predicted probability of septic arthritis was determined for all sixteen combinations of these four predictors and is summarized as less than 0.2 percent for zero predictors, 3.0 percent for one predictor, 40.0 percent for two predictors, 93.1 percent for three predictors, and 99.6 percent for four predictors. The chi-square test for trend and the area under the receiver operating characteristic curve indicated excellent diagnostic performance of this group of multivariate predictors in identifying septic arthritis. CONCLUSIONS: Although several variables differed significantly between the group that had septic arthritis and the group that had transient synovitis, substantial overlap in the intermediate ranges made differentiation difficult on the basis of individual variables alone. However, by combining variables, we were able to construct a set of independent multivariate predictors that, together, had excellent diagnostic performance in differentiating between septic arthritis and transient synovitis of the hip in children.

Acute Disease↗
PubMed1995

Intraocular lens power calculation with an improved anterior chamber depth prediction algorithm.

The accuracy of intraocular lens (IOL) power calculation was evaluated in a multicenter study of 822 IOL implantations using the Binkhorst II, Sanders/Retzlaff/Kraff (SRK I, SRK II, SRK/T), Holladay, and Olsen formulas. All but the first of these were optimized in retrospect with calculation of the SRK A-constant, the Holladay surgeon factor, and the Olsen pseudophakic anterior chamber depth (ACD) for each lens style. The ACD prediction of the Olsen formula was based on a previously described regression formula incorporating preoperative ACD, corneal height, axial length, and lens thickness. Among the optical IOL power calculation formulas, the highest IOL power prediction error was found with Binkhorst's and the lowest with Olsen's, which was more accurate than the SRK/T and the Holladay formulas (P < .05). The SRK/T formula was significantly more accurate than the original SRK regression formulas (P < .001). When analyzed for axial length dependence, all formulas showed the least error in the normal range. Error of the Olsen formula was lower than that of the others in the axial length interval 20 mm to 26 mm. No differences in accuracy were found between the optical IOL calculation formulas in eyes with an axial length above 26 mm (P < .05). The accuracy of IOL power calculation can be improved with optical formulas using newer-generation ACD-prediction algorithms.

Adult↗
PubMedDate not supplied

Immune reactivity against a novel HLA-A3-restricted influenza virus peptide identified by predictive algorithms and interferon-gamma quantitative PCR.

The use of appropriate antigenic peptides for the most common human major histocompatibility complex (MHC) alleles is required for the amplification of the autologous cytotoxic compartment and the development of cytotoxic T cell-mediated immunity. The human A2 allele of the MHC plays an important role for the identification of peptide-specific cytotoxic T cells (CTL) against tumor and viral epitopes. Computer-based prediction algorithms, which are available on the Internet, have already proved to be applicable for the identification of novel CTL epitopes. Using the bioinformatics approach, the authors have identified the novel influenza matrix protein-derived and HLA-A3-restricted 9-mer peptide RLEDVFAGK capable of inducing peptide specific CTL reactivity. Peripheral blood mononuclear cells (PBMC) from healthy individuals and patients with lung cancer were pulsed with this peptide and with the well-characterized HLA-A2-restricted influenza A virus matrix peptide GILGFVFTL. Using quantitative PCR (TaqMan; Applied Biosystems, Foster City, CA, U.S.A), reactivity for both peptides was determined by measuring the change in type 1 cytokine (IFN-gamma) expression upon in vitro stimulation. Peptide-specific reactivity matched well with the subsequently determined MHC-class I alleles of the tested individuals. Results from this study indicate that the use of bioinformatics and the PCR-based screening system for the monitoring of T cell reactivity may allow for the identification of novel CTL epitopes.

Algorithms↗
PubMed2004

Limitation of Potts and Guy's model and a predictive algorithm for skin permeability including the effects of hydrogen-bond on diffusivity.

The Potts and Guy's model for skin permeability, log P = alpha log K - beta MV + delta where P is the permeability coefficient of a compound from aqueous solution through human skin in vitro, K and MV are octanol-water partition coefficient and molecular volume of the compound respectively, and alpha, beta, delta are constants, is examined for a data set of 53 miscellaneous compounds. The model will result in over-estimation for penetrants having higher hydrogen-bond donor activity and underestimation for penetrants having no hydrogen-bond donor. A predictive algorithm for skin permeability including the effects of hydrogen-bond on diffusivity is proposed: log P = alpha log K - beta MV - gamma Hb + delta where Hb is the descriptor of hydrogen-bonding capacity of penetrants and gamma is a constant. The calculated log P values from the latter model are in good accordance with respective experimental ones for the data set.

Algorithms↗
PubMed2000

Use of two predictive algorithms of the world wide web for the identification of tumor-reactive T-cell epitopes.

Tumor cells can be effectively recognized and eliminated by CTLs. One approach for the development of CTL-based cancer immunotherapy for solid tumors requires the use of the appropriate immunogenic peptide epitopes that are derived from defined tumor-associated antigens. Because CTL peptide epitopes are restricted to specific MHC alleles, to design immune therapies for the general population it is necessary to identify epitopes for the most commonly found human MHC alleles. The identification of such epitopes has been based on MHC-peptide-binding assays that are costly and labor-intensive. We report here the use of two computer-based prediction algorithms, which are readily available in the public domain (Internet), to identify HL4-B7-restricted CTL epitopes for carcinoembryonic antigen (CEA). These algorithms identified three candidate peptides that we studied for their capacity to induce CTL responses in vitro using lymphocytes from HLA-B7+ normal blood donors. The results show that one of these peptides, CEA9(632) (IPQQHTQVL) was efficient in the induction of primary CTL responses when dendritic cells were used as antigen-presenting cells. These CTLs were efficient in killing tumor cells that express HLA-B7 and produce CEA. The identification of this HLA-B7-restricted CTL epitope will be useful for the design of ethnically unbiased, widely applicable immunotherapies for common solid epithelial tumors expressing CEA. Moreover, our strategy of identifying MHC class I-restricted CTL epitopes without the need of peptide/HLA-binding assays provides a convenient and cost-saving alternative approach to previous methods.

Algorithms↗
PubMed2006

Anopheles gambiae genome reannotation through synthesis of ab initio and comparative gene prediction algorithms.

BACKGROUND: Complete genome annotation is a necessary tool as Anopheles gambiae researchers probe the biology of this potent malaria vector. RESULTS: We reannotate the A. gambiae genome by synthesizing comparative and ab initio sets of predicted coding sequences (CDSs) into a single set using an exon-gene-union algorithm followed by an open-reading-frame-selection algorithm. The reannotation predicts 20,970 CDSs supported by at least two lines of evidence, and it lowers the proportion of CDSs lacking start and/or stop codons to only approximately 4%. The reannotated CDS set includes a set of 4,681 novel CDSs not represented in the Ensembl annotation but with EST support, and another set of 4,031 Ensembl-supported genes that undergo major structural and, therefore, probably functional changes in the reannotated set. The quality and accuracy of the reannotation was assessed by comparison with end sequences from 20,249 full-length cDNA clones, and evaluation of mass spectrometry peptide hit rates from an A. gambiae shotgun proteomic dataset confirms that the reannotated CDSs offer a high quality protein database for proteomics. We provide a functional proteomics annotation, ReAnoXcel, obtained by analysis of the new CDSs through the AnoXcel pipeline, which allows functional comparisons of the CDS sets within the same bioinformatic platform. CDS data are available for download. CONCLUSION: Comprehensive A. gambiae genome reannotation is achieved through a combination of comparative and ab initio gene prediction algorithms.

Algorithms↗
PubMed1996

Prospective evaluation of a dose prediction algorithm for intravenous tobramycin in adolescent and adult patients with cystic fibrosis.

The predictive performance of a new algorithm to calculate the initial daily dose of tobramycin in patients with cystic fibrosis (CF) was prospectively evaluated. Twenty-six patients with CF (15 men, 11 women, 18-45 years of age) with an acute exacerbation of their chronic pulmonary infection were treated with intravenous tobramycin. The initial dose was calculated with a previously presented algorithm. This algorithm was derived from correlation analysis performed on the adjusted daily dose guided by the determination of serum concentrations: dose (mg three times daily) = 90 + 2.13 x LBM (kg), where LBM (male) = (1.1 x body weight [BW]) - (128 x BW2/height2) and LBM (female) = (1.07 x BW) - (148 x BW2/height2). The predictive performance of this algorithm was evaluated comparing the calculated initial daily dose with the adjusted daily dose for peak and trough levels of 9-11 mg/L and 1.0 mg/L, respectively. Mean squared error and mean error were determined as reflections of precision and bias. The predictive performance of the algorithm was compared with historical data on the predictive performance of the standard equation to dose of 3.3 mg/kg body weight three times daily. The dose calculated with the algorithm proved to give peak serum concentrations in a narrower range and to have a greater precision, but bias was equal. Applying the algorithm, more patients had initial peak serum concentrations in the pre-determined range of 9-11 mg/L than when using the standard equation, so fewer dose adjustments had to be made.

Adolescent↗
PubMed1998

Predicting operative risk for coronary artery surgery in the United Kingdom: a comparison of various risk prediction algorithms.

OBJECTIVE: To compare the ability of four risk models to predict operative mortality after coronary artery bypass graft surgery (CABG) in the United Kingdom. DESIGN: Prospective study. SETTING: Two cardiothoracic centres in the United Kingdom. SUBJECTS: 1774 patients having CABG. MAIN OUTCOME MEASURES: Risk factors were recorded for all patients, along with in-hospital mortality. Predicted mortality was derived from the American Society of Thoracic Surgeons (STS) risk program, Ontario Province risk score (PACCN), Parsonnet score, and the UK Society of Cardiothoracic Surgeons risk algorithm. RESULTS: There were significant differences (p < 0.05) between the British and American populations from which the STS risk algorithm was derived with respect to most variables. The observed mortality in the British population was 3.7% (65 of 1774). The mean predicted mortality by STS score, PACCN, Parsonnet score, and UK algorithms were 1.1%, 1.6%, 4.6%, and 4.7% respectively. The overall predictive ability of the models as measured by the area under the receiver operating characteristic curve were 0.64, 0.60, 0.73, and 0.75, respectively. CONCLUSIONS: There are differences between the British and American populations for CABG and the North American algorithms are not useful for predicting mortality in the United Kingdom. The UK Society of Cardiothoracic Surgeons algorithm is the best of the models tested but still only has limited predictive ability. Great care must be exercised when using methods of this type for comparisons of units and surgeons.

Aged↗
PubMedDate not supplied

Performance of a real-time dicrotic notch detection and prediction algorithm in arrhythmic human aortic pressure signals.

A novel algorithm for real-time detection and prediction of the dicrotic notch from aortic pressure waves was evaluated in arrhythmic aortic pressure signals from heart failure patients. A simplified model of the arterial tree was used to calculate real-time aortic flow from aortic pressure. The dicrotic notch was detected at the first negative dip from the calculated flow, prediction of the notch was performed using a percentage of the decreasing flow. The performance of the real-time dicrotic notch detection algorithm (RTDND) was evaluated during severe arrhythmia from aortic pressure signals of 12 patients. The RTDND was able to detect the dicrotic notch in 98.1%. No false positive dicrotic notch identifications were observed. Prediction of the dicrotic notch was tested at 40%, 20%, and 0% of the decreasing calculated aortic flow. The mean time-delays to the notch were 68 +/- 14 ms, 55 +/- 12 ms, and 43 +/- 8 ms, respectively. Given these small variability, intra-beat prediction of the dicrotic notch may be used for real-time intra-aortic balloon counterpulsation inflation timing.

Aged↗
PubMed2006

A stochastic framework for evaluating seizure prediction algorithms using hidden Markov models.

Responsive, implantable stimulation devices to treat epilepsy are now in clinical trials. New evidence suggests that these devices may be more effective when they deliver therapy before seizure onset. Despite years of effort, prospective seizure prediction, which could improve device performance, remains elusive. In large part, this is explained by lack of agreement on a statistical framework for modeling seizure generation and a method for validating algorithm performance. We present a novel stochastic framework based on a three-state hidden Markov model (HMM) (representing interictal, preictal, and seizure states) with the feature that periods of increased seizure probability can transition back to the interictal state. This notion reflects clinical experience and may enhance interpretation of published seizure prediction studies. Our model accommodates clipped EEG segments and formalizes intuitive notions regarding statistical validation. We derive equations for type I and type II errors as a function of the number of seizures, duration of interictal data, and prediction horizon length and we demonstrate the model's utility with a novel seizure detection algorithm that appeared to predicted seizure onset. We propose this framework as a vital tool for designing and validating prediction algorithms and for facilitating collaborative research in this area.

Algorithms↗
PubMed2006

Bayesian data mining of protein domains gives an efficient predictive algorithm and new insight.

Identification of structural domains in uncharacterized protein sequences is important in the prediction of protein tertiary folds and functional sites, and hence in designing biologically active molecules. We present a new predictive computational method of classifying a protein into single, two continuous or two discontinuous domains using Bayesian Data Mining. The algorithm requires only the primary sequence and computer-predicted secondary structure. It incorporates correlation patterns between certain 3-dimensional motifs and some local helical folds found conserved in the vicinity of protein domains with high statistical confidence. The prediction of domain-class by this computationally simple and fast method shows good accuracy of prediction-average accuracies 83.3% for single domain, 60% for two continuous and 65.7% for two discontinuous domain proteins. Experiments on the large validation sample show its performance to be significantly better than that of DGS and DomSSEA. Computations of Bayesian probabilities show important features in terms of correlation of certain conserved patterns of secondary folds and tertiary motifs and give new insight. Applications for improved accuracy of predicting domain boundary points relevant to protein structural and functional modeling are also highlighted.

Algorithms↗
PubMed2006

Peptide binding motif predictive algorithms correspond with experimental binding of leukemia vaccine candidate peptides to HLA-A*0201 molecules.

The ability to reliably identify the peptides that can bind to MHC molecules is of practical importance for rapid vaccine development. Several computer-based prediction methods have been applied to study the interaction of MHC class I/peptide binding. Here we have compared the binding of peptides predicted by three algorithms (BIMAS, SYFPEITHI and Rankpep) to the binding of the peptides to HLA-A*0201 molecules in vitro, assessed using a MHC stabilization assay on live T2 cells. Fifty HLA-A*0201 peptides were selected from several target oncoproteins: Wilms' tumor protein (WT1), native and imatinib-mutated bcr-abl p210, JAK2 protein and Ewing's sarcoma fusion protein type 1. The sensitivity and specificity of BIMAS, SYFPEITHI and Rankpep respectively, were: 86%, and 82%; 75% and 73%; 64% and 82%. Combining two or more computer methods did not appear to significantly improve the predictive value.

Algorithms↗
PubMed2003

Reliability measures for membrane protein topology prediction algorithms.

We have developed reliability scores for five widely used membrane protein topology prediction methods, and have applied them both on a test set of 92 bacterial plasma membrane proteins with experimentally determined topologies and on all predicted helix bundle membrane proteins in three fully sequenced genomes: Escherichia coli, Saccharomyces cerevisiae and Caenorhabditis elegans. We show that the reliability scores work well for the TMHMM and MEMSAT methods, and that they allow the probability that the predicted topology is correct to be estimated for any protein. We further show that the available test set is biased towards high-scoring proteins when compared to the genome-wide data sets, and provide estimates for the expected prediction accuracy of TMHMM across the three genomes. Finally, we show that the performance of TMHMM is considerably better when limited experimental information (such as the in/out location of a protein's C terminus) is available, and estimate that at least ten percentage points in overall accuracy in whole-genome predictions can be gained in this way.

Algorithms↗
PubMed2002

Prospective trial of a predictive algorithm to transplant cadaver kidneys into highly sensitized patients.

BACKGROUND: The difficulty of transplanting sensitized patients increases proportionally to the panel reactive antibody (PRA) titer. Because of the high likelihood of a positive final crossmatch, these patients are often excluded from a prospective transplant unless there is a 0 HLA-A, -B, -DR mismatch. To address this problem, we developed a computerized algorithm, termed the Kentucky Antibody Testing System (KATS), that predicts class I HLA antigens that would be both "unacceptable" and "acceptable" to the recipient. This report describes the results of a prospective trial among voluntarily participating centers that agreed to share kidneys based on the KATS predictions for patients whose PRA exceeded 40%. METHODS: The results of three antibody screens on each patient were compared with the HLA phenotypes of the cells in the panel in 2x2 tables with calculation of chi-square and correlation coefficient statistics. Private, broad, and public antigens were identified and a list of acceptable and unacceptable antigens were entered into the UNOS computer for each patient listed in the KATS sharing algorithm. RESULTS: Of the total 418 patients meeting the inclusion/exclusion criteria, the largest single group submitted was Black-not-of-Hispanic-origin females. The mean PRA of the patients was 72%. The first transplant via KATS allocation was performed on March 8, 1997. Between that time and the last transplant on July 31, 2000, 145 kidneys were offered to the participating centers and 48 transplants were performed. Of the many reasons listed for not accepting an offer or not transplanting the shared kidney into its intended recipient, only two occurred because of a positive T cell crossmatch and six because of a positive B cell crossmatch. As compared to all other high PRA patients within Southeastern Organ Procurement Foundation who were transplanted during the study period, they were more likely to be non-Caucasian, to be less well matched for private HLA-A, B, and DR antigens, and to have waited for a longer time than the other groups. Although there was a higher incidence of delayed graft function, there was no significant difference in cold ischemia, rejection episodes, or patient or graft survival. CONCLUSIONS: We conclude that KATS, or some other system to prospectively identify a list of acceptable and unacceptable HLA antigens, could improve the access of highly sensitized patients to a successful kidney transplant.

Adult↗
PubMed2006

Rank-statistics based enrichment-site prediction algorithm developed for chromatin immunoprecipitation on chip experiments.

BACKGROUND: High density oligonucleotide tiling arrays are an effective and powerful platform for conducting unbiased genome-wide studies. The ab initio probe selection method employed in tiling arrays is unbiased, and thus ensures consistent sampling across coding and non-coding regions of the genome. Tiling arrays are increasingly used in chromatin immunoprecipitation (IP) experiments (ChIP on chip). ChIP on chip facilitates the generation of genome-wide maps of in-vivo interactions between DNA-associated proteins including transcription factors and DNA. Analysis of the hybridization of an immunoprecipitated sample to a tiling array facilitates the identification of ChIP-enriched segments of the genome. These enriched segments are putative targets of antibody assayable regulatory elements. The enrichment response is not ubiquitous across the genome. Typically 5 to 10% of tiled probes manifest some significant enrichment. Depending upon the factor being studied, this response can drop to less than 1%. The detection and assessment of significance for interactions that emanate from non-canonical and/or un-annotated regions of the genome is especially challenging. This is the motivation behind the proposed algorithm. RESULTS: We have proposed a novel rank and replicate statistics-based methodology for identifying and ascribing statistical confidence to regions of ChIP-enrichment. The algorithm is optimized for identification of sites that manifest low levels of enrichment but are true positives, as validated by alternative biochemical experiments. Although the method is described here in the context of ChIP on chip experiments, it can be generalized to any treatment-control experimental design. The results of the algorithm show a high degree of concordance with independent biochemical validation methods. The sensitivity and specificity of the algorithm have been characterized via quantitative PCR and independent computational approaches. CONCLUSION: The algorithm ranks all enrichment sites based on their intra-replicate ranks and inter-replicate rank consistency. Following the ranking, the method allows segmentation of sites based on a meta p-value, a composite array signal enrichment criterion, or a composite of these two measures. The sensitivities obtained subsequent to the segmentation of data using a meta p-value of 10-5, an array signal enrichment of 0.2 and a composite of these two values are 88%, 87% and 95%, respectively.

Algorithms↗
PubMed2004

HYPROSP: a hybrid protein secondary structure prediction algorithm--a knowledge-based approach.

We develop a knowledge-based approach (called PROSP) for protein secondary structure prediction. The knowledge base contains small peptide fragments together with their secondary structural information. A quantitative measure M, called match rate, is defined to measure the amount of structural information that a target protein can extract from the knowledge base. Our experimental results show that proteins with a higher match rate will likely be predicted more accurately based on PROSP. That is, there is roughly a monotone correlation between the prediction accuracy and the amount of structure matching with the knowledge base. To fully utilize the strength of our knowledge base, a hybrid prediction method is proposed as follows: if the match rate of a target protein is at least 80%, we use the extracted information to make the prediction; otherwise, we adopt a popular machine-learning approach. This comprises our hybrid protein structure prediction (HYPROSP) approach. We use the DSSP and EVA data as our datasets and PSIPRED as our underlying machine-learning algorithm. For target proteins with match rate at least 80%, the average Q3 of PROSP is 3.96 and 7.2 better than that of PSIPRED on DSSP and EVA data, respectively.

Algorithms↗